86 



D. E. ALLEN 



the very small range (only one millimetre) that is involved. H. Stuart Thompson, in a note 

 in his herbarium, also came to the conclusion that every form occurs between the type 

 and var. longidentata, but he failed to reahse that this intergradation follows a definite 

 pattern. The manner, in fact, in which one extreme gradually gives way to the other 

 following a broadly recognizable geographical trend suggests that a cline of some kind 

 may be distinguishable. 



In the earlier determinations of herbarium material during the present study, the 

 two extremes were treated as subspecies - a view which later had to be abandoned. The 

 case for subspecific rank rests not only on the very distinct geographical range of var. 

 longidentata, but also on a number of other characters which various authors have claimed to 

 be associated with the elongation of the calyx teeth. Thus, Knerskon (1874) 

 has asserted that the variety is also characterised by sessile (rather than subsessile) flowers 

 and longer, subherbaceous (rather than scarious) bracteoles. Battandier & Trabut (1890) 

 mention a strongly papillose stigma and sublinear (not ovoid) placentas as additional 

 features. It has not been possible to check these alleged correlations (which would have 

 to be done by a biometric analysis, preferably of fresh material), and it is by no means 

 out of the question for the gene or genes determining the elongated character of the calyx 

 teeth to possess a pleiotropic effect, or even for linkage to be involved. However, since 

 the variation between the two extremes of the calyx character evidently forms a single 

 continuum, the maintenance of two distinct subspecies seems inappropriate. 



It should be stressed that while var. portula, with almost obsolete calyx teeth, is a 

 well-defined extreme, the boundary between var. longidentata and the various intermediate 

 grades is a somewhat arbitrary one. Even so, by studying the variation in an area such as 

 Cornwall, where only two out of the eleven gatherings seen fail to exhibit extreme elonga- 

 tion of the calyx teeth, the limits of var. longidentata can be comprehended and defined 

 with reasonable precision. Moreover, there are specimens of the variety in Gay's herbarium 

 which he records having grown from seed in 1845. In the British Isles it is not uncommon 

 to find both var. portula and intermediates growing together in abundance without any 

 sign of var. longidentata ; I have personally encountered this in populations in Dorset 

 and the Isle of Man, and it is reflected in a number of herbarium gatherings. It was this 

 phenomenon, no doubt, that led Druce to the belief that the true variety of the Iberian 

 Peninsula was not to be found in Britain. 



In the meantime, it seems best to treat these two extreme forms as varieties ; those 

 who like to use a name for the medial part of a continuous series might find var. dentata 

 Druce convenient for the purpose. 



Distribution 



A study of the distribution of the variants reveals a fairly well-marked geographical 

 trend : var. longidentata is prevalent in the west, but becomes gradually replaced eastwards, 

 first by specimens exhibiting various degrees of intermediacy and then by the well-marked 

 extreme of var. portula. It is possible that climate operates effectively in producing a 

 balance between the varieties, and for this reason it is of interest to see how far the isophenes 

 coincide with any particular climatic feature. 



In order to assess the degree of oceanicity of the climate, Greig-Smith (1950) has 

 employed Amann's Index of Hygrothermy. This is obtained by dividing the annual 

 precipitation (in centimetres) multiplied by the annual mean temperature by the mean 

 temperature of the coldest month subtracted from that of the hottest month, the values 

 of these last three being expressed in degrees Centigrade. On comparing the values for 

 this index in various parts of the British Isles with the variation in Peplis portula as plotted 

 on the map (see fig. 1), there would appear to be some correlation in Great Britain between 



